What a good origination pilot actually looks like.

Almost every origination provider will offer you a pilot. Almost none of them will tell you what makes a pilot worth running versus a paid trial designed to fail quietly. This is a checklist, not a pitch: what to look for before you agree to one, whether it is with us or anyone else.

01A real, narrow, testable thesis

A pilot cannot test "explore the market broadly." It can test one specific thesis: a sector, a size band, a geography, an ownership profile. If the brief going into the pilot is broad, the pilot will produce a broad, unfalsifiable result, and you will learn nothing about whether the approach actually works for what you buy. Narrow the thesis before the clock starts, not during it.

02A stated evaluation window, not an open-ended trial

A good pilot has an agreed length before it begins, long enough for the real mechanics to run: the market gets mapped, the fit model scores the universe, trigger detection has time to surface real signals, and outreach has time to land and get a reply. Across engagements we have run, the earliest real signal typically appears within the first few weeks, not the first few days. A pilot too short to reach that point tests nothing except how fast a provider can send a first email.

03"Qualified" is defined in writing before it starts

This is where most pilots quietly fail. If "qualified conversation" is not defined and agreed before outreach begins, both sides grade the result differently at the end, and the disagreement becomes about the definition instead of the outcome. Agree it up front: does a qualified conversation mean the owner replied, replied with genuine interest, or agreed to a call. Write it down. A provider unwilling to fix this definition before you start is telling you something.

A vague brief produces vague output. The same is true of a vague pilot. The fix is the same: write down what counts, before either side has an incentive to move the goalposts.

04Real deliverables during the window, not just a promise

You should be able to see the work happening, not just wait for a final report. A mapped target universe you can inspect. Real founder conversations, with names and dates, as they happen. A running count against the definition agreed in step three. If the only visible output during the pilot is a dashboard of "activity" with no conversations to show for it, the pilot is measuring effort, not origination.

05A capacity check on your side too

Origination fails on the receiving end as often as the sending end. Before the pilot starts, confirm someone on your team can take an owner call within a few days of them raising a hand, for the full length of the window. A pilot that produces five genuine conversations and loses three of them to slow follow-up on your side will look like the engine failed, when the actual failure was capacity. See who this is not a fit for if that capacity does not exist yet.

06A clear decision point at the end

A good pilot ends with a real decision, not an automatic renewal. Three honest outcomes: the thesis was right and the volume justifies continuing, the thesis needs adjusting and the pilot bought you the evidence to adjust it, or the approach is not the right fit and you walk away having spent a defined, bounded amount to find that out. Any of the three is a good result. What is not a good result is a pilot structured so vaguely that nobody can tell which of the three actually happened.

07What we will actually agree with you

We will not publish a fixed pilot length, fee, or deliverable count here, because the honest answer depends on your thesis, your universe size, and your capacity to take conversations, which are exactly the things this checklist asks you to nail down first. What we will do on a call is work through all six points above against your specific mandate and agree a structure in writing before anything starts, using this checklist as the actual agenda.